As Retraction Watch readers know, criminal sanctions for research fraud are extremely rare. There have been just a handful of cases — Dong-Pyou Han, Eric Poehlman, and Scott Reuben, to name several — that have led to prison sentences.
According to a new study, however, the rarity of such cases is out of sync with with the wishes of the U.S. population:
[T]he public overwhelming judges both data fraud and selective reporting as morally wrong, and supports a range of serious sanctions for these behaviors. Most notably, the vast majority of Americans support criminalizing data fraud, and many also believe the offense deserves a sentence of incarceration.
The new study was written by Justin Pickett and Sean Patrick Roche, both of the University at Albany (SUNY). The paper, “Public Attitudes Toward Data Fraud And Selective Reporting in Science,” was posted to the SSRN preprint server, and has been submitted to a peer-reviewed journal, Pickett told Retraction Watch. He added that he and Roche are happy to share or post the data for anyone who wants to analyze it.
Pickett and Roche conducted two surveys, one of more than 800 people in what is known as a convenience sample, of those easy to reach (in this case, using Amazon’s Mechanical Turk), and another of a representative sample of more than 950 people in the U.S. The first survey showed that
There is an extraordinary consensus among participants that both falsifying or fabricating data and selective reporting are morally unacceptable.
Respondents viewed falsification, fabrication, and selective reporting somewhat differently, but not dramatically so:
Over 90% of participants believe that scientists caught falsifying or fabricating data should be fired and banned from receiving government funding. However, most participants also believe that selective reporting deserves these same sanctions. The majority of participants believe that data fraud should be a criminal offense. Well over a third of participants hold the same view of selective reporting.
[P]articipants absolutely do not view selective reporting as a “questionable research practice”; rather, the vast majority of laypersons in our sample believe this common behavior is morally reprehensible.
In the second survey — the representative sample — respondents felt even more strongly:
In the general population sample, most respondents who support criminalization prefer a sentence of incarceration, rather than a fine and/or probation. The results indicate that slightly over half of all Americans would prefer both to criminalize data fraud and to sentence fraudsters to a period of incarceration.
In other words:
The American public thus appears to be very punitive toward fraudulent scientific behavior, supporting a much harsher punishment approach than is currently in use.
(Incidentally, our recent informal poll produced similar results — 8 out of 10 respondents said scientific fraud should be prosecuted as a crime.)
Pickett answered a few questions about the study for us.
RW: Are there any limitations in the methodology you used, that would affect the results, or make them less generalizable?
Pickett: The main limitations are as follows. First, the sample in Study #1 is a convenience sample (from MTurk), which means there is no basis (from probability theory) for assuming generalization. However, the sample in Study #2 is a national probability sample, which does not have this limitation. That study, however, has a low response rate―29% or 2%, depending on how you define it. At the same time, meta-analytic research shows that response rates are not predictive of nonresponse bias (Groves, 2006; Groves and Peytcheva, 2008). A recent report to the National Science Foundation emphasizes this point: “nonresponse bias is rarely notably related to [the] nonresponse rate” (Krosnick et al., 2015).
One limitation is that we may have obtained different results if we had provided respondents with different examples of fraud and selective reporting. For example, we didn’t provide any information about funding in the examples. We tried to use realistic examples, and we made sure that the examples pertained to different scientific disciplines. Nonetheless, I am sure that if we had focused on cancer studies and told them the studies are often government-funded, we would have gotten even higher levels of support for criminalization.
RW: One of the arguments we often hear against criminalizing scientific fraud is that it would make whistleblowers think twice about bringing their concerns forward. Do these findings bear on that potential concern?
Pickett: …I’m not sure if the findings directly bear on the concern about whistleblowers. Is there actually any evidence that criminalizing fraud would have this effect, or is it just speculation? I would actually think criminalization would have the opposite effect. Many expressive theories of law suggest that criminalization and punishment sends an important signal about the moral wrongfulness of behavior and can actually increase the strength of prohibitive norms. I would think that stronger norms against these behaviors would lead to greater reporting, not less.
RW: Others have pointed out that when it wants to, the Department of Justice can prosecute, once the Office of Research Integrity has referred a case. But that happens extremely rarely. What do these results suggest about the frequency with which this occurs?
Pickett: [T]he results suggest the public would probably want the cases to be referred and prosecuted much more frequently.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.